Forest Ecology And Management - Harvard University
Forest Ecology and Management 255 (2008) 4021–4031 Contents lists available at ScienceDirect Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco The effect of logging on vegetation composition in Western Massachusetts Robert I. McDonald a,*, Glenn Motzkin b, David R. Foster b a b The Nature Conservancy, Worldwide Office, 4245 North Fairfax Drive, Arlington, VA 22203, United States Harvard Forest, Harvard University, 324 North Main Street, Petersham, MA 01366, United States A R T I C L E I N F O A B S T R A C T Article history: Received 6 December 2007 Received in revised form 19 February 2008 Accepted 26 March 2008 Forest harvesting is one of the most significant disturbances affecting forest plant composition and structure in eastern North American forests, yet few studies have quantified the landscape-scale effects of widespread, low-intensity harvests by non-industrial private forest owners. Using spatially explicit data on all harvests over the last 20 years, we sampled the vegetation at 126 sites throughout central and western Massachusetts, one-third of which had not been harvested, and two-thirds of which had been harvested once since 1984. Seedling and sapling densities increased with increasing harvest intensity, but decreased to levels similar to unharvested sites by year 20 for all but the most intensive harvests. The composition of understory trees appears to be only slightly changed by harvesting, and was strongly correlated with adult tree composition. Regeneration was dominated by Betula lenta followed by Pinus strobus; Quercus spp. exhibited little sapling recruitment, even in Quercus-dominated stands. Total vascular plant species richness increased substantially with harvesting on low C:N sites (i.e., rich soils), but was only slightly increased on high C:N sites. While harvesting was associated with a statistically significant change in vascular plant composition, non-metric multidimensional scaling revealed that climate (temperature, precipitation) and C:N ratios were the major correlates of composition. Overall, the compositional impacts of harvesting were minor, perhaps because of the low-intensity of harvesting. However, our results support observations from elsewhere in the northeastern U.S. of limited oak regeneration on both harvested and unharvested sites. In addition, our results suggest that increased harvest intensity may be expected to alter forest composition, particularly on rich sites where invasive species may increase as a result of harvesting. ß 2008 Elsevier B.V. All rights reserved. Keywords: Indicator species analysis (ISA) Land-use legacy Multi-response permutation procedures (MRPP) Pinus strobus Quercus Selective cutting 1. Introduction Southern New England, like much of the eastern United States, is characterized by extensive non-industrial private forests (NIPF). Individual forest landowners in the region generally own small properties, whose size has decreased over time as the landscape has become more parcelized (average 5 ha; Kittredge, 1996; Kittredge et al., 2003). Small parcel size, combined with landowner preferences that often do not prioritize income-generating activities from the land (Finley and Kittredge, 2006), has led to a pattern of relatively frequent, dispersed, low-intensity harvests. For instance, in western and central Massachusetts, the study area of this paper, about 1% of forest area was harvested each year over the past two decades. However, the average harvest volume was 43 m3/ha, which is about one-fifth of typical total stand volume (McDonald et al., 2006). We conducted field studies to determine * Corresponding author at: 4245 North Fairfax Drive, Arlington, VA 22203, United States. Tel.: 1 703 841 5300. E-mail address: rob mcdonald@TNC.ORG (R.I. McDonald). 0378-1127/ – see front matter ß 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2008.03.054 the effects of this low-intensity harvesting regime, attempting to answer two important questions. First, there is regional concern about regeneration of moderately shade-tolerant tree species. In particular, although oaks (Quercus spp.) dominate extensive areas today, they appear to regenerate infrequently, mirroring a trend towards declining oak throughout the eastern U.S. (Abrams, 2003). Increased forest harvesting has been proposed as one solution to increase oak regeneration (e.g., Bellocq et al., 2005). However, it is unclear whether the low-intensity harvests typical of NIPF owners will have this effect. Low-intensity harvests may not dramatically alter the environment at the forest floor, and may merely release from suppression existing stems in the understory, especially of more shade-tolerant species, like red maple (Acer rubrum; cf., Abrams, 1998). In contrast, high-intensity harvests may allow intolerant and moderately shade tolerant species, including oaks, to regenerate in response to changes in light, temperature, and moisture and increased soil scarification (cf., Chen et al., 1993; Saunders et al., 1999; Davies-Colley et al., 2000; Newmark, 2001; Gray et al., 2002). We assess the effect of low-intensity harvests on the regeneration of two dominant taxa (Pinus strobus and Quercus
4022 R.I. McDonald et al. / Forest Ecology and Management 255 (2008) 4021–4031 rubra), evaluating time since harvest in order to determine whether increased seedling establishment immediately after a cut results in increased abundance of saplings over time. Second, we evaluate whether harvesting significantly alters the composition and richness of the vascular vegetation in a landscape characterized by frequent, low-intensity cuts. Although this topic has been much discussed in the literature, results of stand-level studies have differed substantially (Gilliam, 2002), and it is not clear a priori whether harvesting will increase or decrease species richness across a varied landscape. Immediately after a harvesting event, a set of disturbance-tolerant species often thrives; these species may subsequently decline with increasing time since disturbance. In addition, a set of forest interior species may decrease locally because of direct harvesting impacts, altered resource conditions, increased competition in the post-harvest environment, or slow rates of reestablishment and growth (Meier et al., 1995; Roberts and Gilliam, 1995). In this analysis, we track how net vascular species richness changes with time since harvest, and evaluate whether harvesting effects vary across a gradient of soil fertility. As is often the case when working at a landscape-scale, other confounding factors must be considered in such an analysis. Throughout the study area, there is a moderate gradient in temperature and precipitation, with higher elevations being cooler and wetter. Similarly, soil texture and nutrient composition vary spatially and can substantially affect species composition. Of particular importance in the region are small areas of calcareous or mafic bedrock characterized by rich soils with high base saturation, which often have distinctive floras (Bellemare et al., 2005). The legacy of past agricultural disturbance is also an important control of modern vegetation composition across the region (Foster and Aber, 2004). Southern New England was heavily forested prior to the arrival of European settlers, although Native Americans may have had significant effects on local forest composition and structure (cf., Cronon, 1983). By the mid-19th century, more than 60% of the region’s forests were cleared for agriculture, leaving a patchwork of plowed fields, pastures, and active woodlots (Hall et al., 2002). Widespread agricultural abandonment by the early decades of the 20th century allowed most of these forests to re-grow, albeit with altered vegetation composition (e.g., Foster et al., 1998, 2003; Motzkin et al., 1999). In this paper, we assess potential effects of past land-use on soil characteristics and plant composition, while focusing our analysis on the effects of recent harvesting. Specific questions addressed in this study include: (1) What affect does recent forest harvesting have on the composition of tree regeneration and has forest harvesting resulted in increased regeneration of commercially valuable species (especially Q. rubra Fig. 1. Environmental variation and location of study sites in central and western Massachusetts. Upper left: Forested areas in 1999 that were unharvested over the period 1984–2003 (shown in green), versus forests harvested at various intensities. White areas were not forested in 1999. Upper right: Bedrock geology (from MassGIS). Lower left: Forest cover in the 1830s (Hall et al., 2002). Some towns do not have historical data available. Lower right: The distribution of sampled sites across the study area. Grey lines show the boundaries of the physiographic regions referred to in this paper (Motts and O’Brien, 1981).
R.I. McDonald et al. / Forest Ecology and Management 255 (2008) 4021–4031 and P. strobus)? (2) Does forest harvesting result in increased or decreased vascular plant species richness, and do such changes persist over time? (3) Does the influence of forest harvesting on species richness and composition vary depending on edaphic conditions? 2. Materials and methods 2.1. Site description Our study area is the forested portions of Massachusetts west of the coastal lowlands (Fig. 1), as defined by Motts and O’Brien (1981). Bedrock is primarily granite, gneiss, and schist, with areas of calcareous and mafic bedrock in the western portion of the study area. The dominant soil parent material is glacial till, with alluvial deposits in the Connecticut and Berkshire Valleys. Soils for much of the region are Inceptisols, with valley floodplains dominated by Entisols, and the Western Upland (Fig. 1) having a mix of Inceptisols and Spodisols (Brady and Weil, 2002). The number of growing degree days varies from more than 3800 in the Connecticut River Valley (8C, with 0 8C baseline) to less than 2700 in high elevation sites in the Western Uplands. Precipitation varies along a similar gradient, being greatest in the Western Upland region (147 cm year 1) and lowest in the Connecticut River Valley (97 cm year 1), with snowfall being more prevalent at higher elevations (National Climatic Data Center, 2006). 2.2. Harvesting patterns Massachusetts requires all commercial harvest operations above 87 m3 volume to file a Forest Cutting Plan (FCP), which includes a map of the harvest area and information on species, harvest volumes, and silvicultural objectives. We digitized all FCP harvests from 1984 to 2003 (McDonald et al., 2006). Across the study area, harvesting is most frequent in the Western and Central Uplands (1.3% forest harvested annually) and least frequent in the Connecticut and Berkshire Valleys (0.8% forest harvested annually). Average reported harvesting volume is remarkably consistent over the study area (overall mean 43 m3/ha; McDonald et al., 2006). Previous research has found that, prior to harvest, logged sites are similar to unharvested sites with respect to site conditions, species composition, and structure (Kittredge et al., 2003). 2.3. Data collection We selected 126 sites for sampling (Fig. 1). Two-thirds of the sites (N 89), selected randomly from the larger FCP database, were harvested once between 1984 and 2003. Harvest volume varied from 3 m3/ha to 231 m3/ha, with a mean of 43 m3/ha. The remaining one-third of the sites (N 37) were control sites selected randomly from all forest areas that were unharvested since at least 1984. Sampled sites were predominately on non-industrial private forest land (N 105), with some state (N 15), municipal (N 5), and industrial (N 1) forest sites. Cutting patterns were similar across ownership types, although harvest intensity was slightly higher on state-owned land near the Quabbin Reservoir (see McDonald et al., 2006 for a detailed analysis). The boundaries for our field surveys were those filed with the FCP for harvested sites, and were circular for control sites, proportional in size to the mean FCP in each physiographic region (Motts and O’Brien, 1981): Connecticut River Valley (10.7 ha), Central Upland (11.3 ha), Berkshire Valley and Taconic Mountains (12.6 ha), and the Western Upland (15.9 ha). For towns where information on 1830s land cover (i.e., open versus forested) was available (Hall et al., 2002), we insured that our field sites were of a 4023 consistent land cover. Note that our sample sites were welldispersed over the study area (Fig. 1), and encompassed the variety of past agricultural land-uses. Field sampling was conducted during the summers of 2004 and 2005. While many vernal species were still present in some form, we may have missed a few ephemeral vernal species. Based on reconnaissance of the entire polygon, our field sampling first determined the distribution of harvesting within the polygon, identified the major stand type (i.e., hardwood, evergreen, or mixed), and recorded the presence and abundance of non-native invasive species (see McDonald et al., in press). Within the major stand type (i.e., excluding small areas of the polygon of different types) we randomly placed 10 plots of 11.3 m radius (400 m2), at least 50 m from one another. Tree basal area (m2/ha) and density (#/ha) were measured for overstory ( 20 cm DBH) and midstory (5–20 cm DBH) trees using the point-quarter method (Cottam and Curtis, 1956). As an estimate of local harvest intensity, the diameter of all stumps from harvested trees within each plot was also recorded. Seedlings (stems less than 1.37 m) and saplings (stems greater than 1.37 m and less than 5 cm DBH) were counted within 2 m and 5 m radius plots, respectively. Taxonomy follows Gleason and Cronquist (1991). In addition, within the dominant stand, a 20 m 20 m intensive plot was randomly located in which we estimated the percent cover of each vascular plant species using an eight-class scale: (1 1%, 2 1–3%, 3 3–5%, 4 5–15%, 5 15–25%, 6 25–50%, 7 50–75%, 8 75%). Given the generally low species richness and herbaceous cover in these plots, we feel confident that all vascular plant species present were recorded. We also measured the basal diameter of all stumps within the plot and in a 5 m buffer around it. A soil pit was dug to 50 cm to examine the soil profile for evidence of a plow layer or other soil disturbance. Using this information, as well as other evidence of land-use such as the presence of stone walls, barbed wire, etc., we classified the site into one of three categories of historical land-use: woodlot, if the site showed no evidence of having been used for historical agriculture and was likely to have been continuously forested; pasture, if the site was used as rough, unimproved pasture; and plowed, if there was clear evidence of an Ap horizon (Motzkin et al., 1999). We then compared these field observations with land-use data from the 1830s (Hall et al., 2002) and from 1930s maps of the Works Progress Administration, where available, to assign each site to historical land-use categories. Based on our field reconnaissance and the historical data sources, we are confident that the land-use history determined for the 20 m 20 m plot is accurate for the majority of the area of each polygon. Two soil samples were taken of the organic soil and the 0–15 cm mineral soils, respectively, to characterize soil texture and nutrients (Brookside Laboratories). Slope, aspect, canopy closure, and percent surface rock were estimated for the plot. A 5 m resolution digital elevation model from MassGIS, re-sampled to 30 m resolution, was used to calculate topographic convergence index (TCI, Beven and Kirkby, 1979), a measure of topographically derived wetness, and to interpolate growing degree days and precipitation from metrological stations, using published equations for the region (Ollinger et al., 1995). 2.4. Data analysis Summary statistics of woody basal area and density were calculated for the overstory and midstory from the 10 plots within the major stands. Stump basal area, recorded at ground level, was calculated in aggregate for each polygon, as well as separately for each of the 10 plots. Vascular plant data from the intensive 20 m 20 m plot were converted to percent cover using the midpoint of the cover classes. For analyses comparing the percent
4024 R.I. McDonald et al. / Forest Ecology and Management 255 (2008) 4021–4031 cover data to harvesting intensity, stump basal area in and around the 20 m 20 m plot was used, rather than stand-wide averages. For descriptive purposes, overstory relative basal area was used to cluster our plots into a small set of discrete groups. All clustering was done using Ward’s method for hierarchical grouping in PCORD 5.05, with dissimilarities calculated using a Euclidean distance metric, as is consistent with the intergroup distance calculation implicit in Ward’s method (McCune and Grace, 2002). The average species composition in the overstory, midstory, sapling, and seedling strata was calculated for each of our compositional groups. Harvesting was categorized into two groups: moderate/ heavy (greater than 10 m2 of basal area of stumps per ha) or light/ none (0–10 m2 of basal area of stumps per ha). Data on species composition were ordinated with non-metric multidimensional scaling (McCune and Grace, 2002). All ordinations were conducted in PCORD, using the Bray-Curtis metric to quantify dissimilarity among sites (McCune and Grace, 2002). For vascular plant cover data, ordinations used presence/absence data for species that occurred in more than 5% of the sites sampled, to avoid bias from rare species. Ordination results using abundance data were similar, and are not presented here. Taxonomic lumping was conducted where necessary to insure consistency among field crews across the field season (see Table 5). For all ordinations, PCORD’s ‘‘slow but thorough’’ autopilot algorithm was used, with 50 Monte Carlo samples. The potentially explanatory variables described above were correlated with the ordination axes to evaluate correlations with species composition. To test the relationships between harvesting, soil fertility, and species composition, we used multi-response permutation tests (MRPP). With MRPP, the statistical question is: Are sites in the same group more similar in species composition than sites in different groups (McCune and Grace, 2002)? Harvesting was categorized into two groups as above: moderate/heavy or light/ none. If the MRPP test was significant, an Indicator Species Analysis (Dufrêne and Legendre, 1997) was used to determine what species have high abundance and frequency in each group. Finally, to compare species composition in the overstory and midstory (combined) with the sapling or seedling layer, we conducted a Mantel test (Mantel, 1967; Smouse et al., 1986) of correspondence between the two strata. To examine the effect of forest harvesting on sapling and seedling densities, as well as species richness in the intensive plots, we used generalized linear models (GLMs). For sapling and seedling densities, a negative binomial error distribution and log link function was appropriate, and model fitting was done with function glm.nb in the MASS library of SPLUS. For species richness a Gaussian error distribution was used. The number of years since a harvest was added to our set of potential explanatory variables, but was allowed to enter the model only as an interaction with harvest intensity (m2/ ha of stumps). For the sapling and seedling analyses, a fixed Polygonlevel factor was added, to account for differences in means among sites. For all three analyses, forward stepwise selection was performed using function step.aic in the MASS library of SPLUS. 3. Results Soil characteristics vary with bedrock-type and prior land-use (Table 1). Percent sand is correlated with prior land-use, with sandier soils in formerly pastured or plowed sites than in continuously wooded sites. Base saturation varies with bedrock type and prior land-use, with greater base saturation on calcareous than granitic bedrock, and greater base saturation in plowed sites than in continuously wooded sites. C:N ratios are lower on sites with calcareous versus granitic bedrock. P and S have no apparent relationship with bedrock-type or prior land-use. Table 1 Soil characteristics for study sites, as a function of bedrock type and prior land-use (BS base saturation) Bedrock type Granitic Mafic Calcareous Prior land-use Woodlot 47.1% 11.1% sand a 46.6% 14.8% sand a 42.9% 5.2% sand a 31.1% 6.4% BSa 34.9 1.1% BSa,b 36.0% 9.0% BSb 34.1 ppm 60.0 ppm Pa 26.3 ppm 12.8 ppm Pa 16.0 ppm 7.7 ppm Pa 18.8 4.7 C:Na 17.9 4.8 C:Na,b 15.7 2.8 C:Nb 43.5 23.0 Sa 34.3 9.9 Sa 49.0 24.8 S a Pasture 52.9% 13.6% sanda,b 54.3% 13.9% sanda,b 39.4% sanda,b 31.7% 5.4% BSa,c 38.7% 10.5% BSa,b,c,d 45.3% 14.1% BSb,d 26.1 ppm 28.6 ppm Pa 34.6 ppm 18.8 ppm Pa 10.0 ppm 1.4 ppm Pa 18.7 3.8 C:Na 18.1 2.1 C:Na,b 13.7 5.3 C:Nb 37.5 14.3 Sa 44.0 14.2 Sa 25.0 10.0 Sa Plowed 52.6% 12.0% sand b 64.8% 16.5% sand b 51.5% 11.3% sandb 37.3 10.5% BSc 35.5% 4.2% BSc,d 48.0% 20.1% BS d 25.4 ppm 20.0 ppm Pa 44.0 ppm 47.6 ppm Pa 28.4 ppm 23.0 ppm Pa 16.8 4.7 C:Na 18.5 0.5 C:Na,b 13.6 3.3 C:Nb 37.8 10.0 Sa 45.8 9.2 Sa 34.4 16.7 S a Superscript letters indicate which groups are statistically identical to one another, and which groups are significantly different. See text for details. Cluster analysis of overstory trees identified four relatively distinct groups: Mixed hardwood, Quercus-dominated, Tsuga canadensis-dominated, and P. strobus-dominated forests. The Mixed hardwood group is characterized by A. rubrum, A. saccharum, and Q. rubra in the overstory (Table 2). Midstory composition is similar, with Q. rubra generally replaced by Fagus grandifolia. Harvested Mixed hardwood sites have lower A. rubrum and Q. rubra basal area, and more A. saccharum, than unharvested sites. On unharvested sites, the seedling layer is dominated by A. rubrum and A. saccharum, and the sapling layer has abundant F. grandifolia. On harvested sites, the seedling layer has substantially more P. strobus and Betula lenta than unharvested sites. The sapling layer has more B. lenta and less F. grandifolia, relative to unharvested sites. The Quercus group is dominated by Q. rubra in the overstory. Midstory composition is a mix of A. rubrum, Q. rubra, and B. lenta. On unharvested sites, the seedling layer is dominated by Q. rubra and A. rubrum, while the sapling layer has abundant B. lenta and A. rubrum. On harvested sites, the seedling layer has more B. lenta, P. strobus, and less F. grandifolia, while the sapling layer has more F. grandifolia, relative to harvested sites. The T. canadensis group is dominated by T. canadensis in the overstory and midstory. On unharvested sites, the seedling layer is dominated by T. canadensis and A. rubrum, and the sapling layer by F. grandifolia and T. canadensis. The seedling layer of harvested sites has more P. strobus and Q. velutina, while the sapling layer has more B. lenta and less T. canadensis, relative to unharvested sites. The P. strobus group is dominated by P. strobus with some Q. rubra. Q. rubra is less abundant in harvested than unharvested sites. Midstory composition is a mix of P. strobus, A. rubrum, B. lenta, and T. canadensis. On unharvested sites, the seedling layer is dominated by P. strobus and A. rubrum, as is the sapling layer. The seedling layer of harvested sites is characterized by more A. rubrum, Q. rubra and B. lenta, and less P. strobus, while the sapling layer has more B. lenta, relative to unharvested sites. Oak regeneration in the oak (and other) forests we sampled is less than that of other species, and harvesting appears to have minor effects on oak regeneration (Fig. 2). Within the Quercus group, there is a trend towards harvested sites (average stump BA: 24.9 m2/ha) having more Quercus seedlings than not/lightly harvested sites (average stump BA: 2.4 m2/ha), but this is not statistically significant (t 0.91, P 0.48). In addition, because
R.I. McDonald et al. / Forest Ecology and Management 255 (2008) 4021–4031 4025 Table 2 Tree composition of the four forest types defined by cluster analysis, by strata and harvesting level Mixed hardwood forest (species) Unharvested Overstory A. rubrum A. saccharum Q. rubra T. canadensis P. strobus Fraxinus americana Fagus grandifolia Other Midstory A. rubrum F. grandifolia A. saccharum T. canadensis Betula lenta Betula alleghaniensis Other Harvested BA relBA Density relDens BA relBA Density 5.4 4.3 4.3 2.2 2.0 2.0 1.5 18.2% 14.5% 14.5% 7.3% 6.7% 6.7% 4.9% 8.2 65 45 34 22 13 21 18 20.3% 14.1% 10.7% 7.0% 4.1% 6.6% 5.5% 2.2 6.4 2.2 1.5 1.4 1.4 1.4 10.1% 29.1% 10.0% 6.8% 6.3% 6.5% 6.6% 25 61 19 15 14 12 16 11.1% 27.6% 8.6% 6.6% 6.2% 5.2% 7.4% 27.3% 101 31.7% 5.4 24.6% 60 27.3% 1.2 0.8 0.7 0.5 0.4 0.3 21.5% 15.2% 12.9% 8.8% 7.7% 4.8% 99 91 67 47 38 26 20.1% 18.5% 13.7% 9.6% 7.7% 5.3% 0.5 0.5 1.2 0.7 0.2 0.3 10.5% 11.0% 25.1% 14.8% 3.5% 6.3% 37 63 129 55 19 27 7.8% 13.3% 27.1% 11.5% 4.0% 5.7% 1.6 29.1% 124 25.0% 1.4 28.8% 146 30.6% 62 111 142 152 391 56 5.0% 8.9% 11.4% 12.2% 31.4% 4.5% 353 345 310 241 207 169 16.3% 15.9% 14.3% 11.1% 9.5% 7.8% 331 26.6% 542 25.0% 378 2479 2248 136 869 1482 955 3.1% 20.0% 18.1% 1.1% 7.0% 12.0% 7.7% 2012 1945 1922 1233 1106 995 901 15.5% 15.0% 14.9% 9.5% 8.5% 7.7% 7.0% 3844 31.0% 2828 21.9% Sapling B. lenta A. saccharum A. rubrum A. pensylvanicum F. grandifolia P. strobus Other Seed P. strobus A. rubrum A. saccharum B. lenta Prunus serotina A. pensylvanicum Q. rubra Other Oak forests (species) Unharvested relDens Harvested relDens BA relBA 170 16 12 14 18 62.6% 6.0% 4.5% 5.1% 6.5% 14.0 1.0 1.3 2.9 0.4 63.4% 4.5% 6.0% 12.9% 1.7% 131 13 13 27 6 58.1% 5.6% 5.9% 12.0% 2.6% 9.9% 42 15.3% 2.6 11.5% 36 15.8% 1.2 0.7 0.6 0.5 0.4 28.6% 16.7% 13.8% 11.3% 10.3% 143 45 58 51 45 33.8% 10.7% 13.9% 12.1% 10.7% 1.1 0.4 0.2 0.3 0.6 30.7% 11.0% 5.6% 9.2% 17.4% 103 23 25 18 58 32.3% 7.3% 8.0% 5.8% 18.0% 0.8 19.2% 79 18.7% 0.9 26.1% 91 28.7% 583 382 27 33 44.7% 29.4% 2.1% 2.5% 621 330 318 183 31.3% 16.6% 16.1% 9.3% Other 277 21.3% 529 26.7% Seed Q. rubra B. lenta 2033 743 24.0% 8.8% 3657 2065 29.7% 16.8% Overstory Q. rubra B. lenta P. strobus T. canadensis A. rubrum Other Midstory A. rubrum Q. rubra B. lenta P. strobus T. canadensis Other Sapling B. lenta A. rubrum F. grandifolia A. pensylvanicum BA relBA 18.1 1.5 1.4 1.3 1.1 69.7% 5.7% 5.3% 5.1% 4.3% 2.6 Density Density relDens
R.I. McDonald et al. / Forest Ecology and Management 255 (2008) 4021–4031 4026 Table 2 (Continued ) Oak forests (species) Unharvested BA Harvested Density relDens P. strobus A. rubrum F. grandifolia 881 2288 32 Other 2484 Hemlock forests (species) Overstory T. canadensis A. rubrum Q. rubra P. strobes B. lenta Other Midstory T. canadensis A. rubrum F. grandifolia B. lenta P. strobes Other relBA relBA Other Sapling B. lenta P. strobus A. rubrum A. saccharum Other Seed A. rubrum 1863 1720 796 15.1% 14.0% 6.5% 29.4% 2202 17.9% Density relDens relBA 16.2 4.1 3.7 2.7 2.2 43.8% 11.1% 10.0% 7.2% 6.1% 159 41 31 21 25 43.3% 11.2% 8.4% 5.8% 6.7% 14.2 3.6 2.9 3.7 1.8 44.7% 11.2% 9.3% 11.5% 5.5% 154 58 31 30 21 42.5% 16.0% 8.5% 8.2% 5.9% 8.0 21.8% 90 24.6% 5.6 17.7% 69 18.9% 2.2 0.7 0.3 0.3 0.3 45.9% 13.4% 6.3% 6.0% 5.3% 190 56 33 29 20 43.3% 12.7% 7.5% 6.7% 4.5% 2.3 1.1 0.1 0.3 0.3 47.8% 22.1% 3.0% 6.5% 5.2% 174 69 12 33 17 46.9% 18.5% 3.3% 8.8% 4.5% 1.1 23.1% 110 25.2% 0.8 15.5% 66 17.8% 70 139 44 230 34 11.6% 23.0% 7.2% 38.1% 5.6% 653 443 350 210 94 35.1% 23.8% 18.8% 11.3% 5.1% 87 14.4% 109 5.9% 3844 1290 252 77 643 610 37.4% 12.5% 2.4% 0.7% 6.3% 5.9% 4749 2862 1742 965 960 826 33.3% 20.1% 12.2% 6.8% 6.7% 5.8% 3569 34.7% 2161 15.1% relDens Other Midstory P. strobes A. rubrum B. lenta T. canadensis A. saccharum Q. velutina 10.4% 27.0% 0.4% BA Seed A. rubrum T. canadensis P. strobes Q. velutina A. pensylvanicum Q. rubra Other relDens relDens Other Overstory P. strobes Q. rubra T. canadensis A. rubrum relBA Harvested BA Sapling B. lenta F. grandifolia A. pensylvanicum T. canadensis P. strobus White pine forests (species) Density Unharvested BA Density Unharvested Harvested BA relBA 18.1 2.9 1.7 1.6 62.5% 10.2% 6.0% 5.7% 4.5 Density relDens BA relBA Density 145 33 20 25 50.8% 11.7% 6.9% 8.7% 12.5 0.9 2.2 2.0 53.0% 4.0% 9.4% 8.5% 98 9 24 30 43.2% 3.8% 10.8% 13.4% 15.6% 62 21.9% 5.9 25.1% 65 28.9% 1.3 0.8 0.5 0.4 0.4 0.2 29.6% 19.0% 11.8% 10.5% 8.8% 5.4% 127 81 52 32 39 20 31.7% 20.1% 12.9% 8.0% 9.8% 4.9% 0.9 0.7 0.4 0.4 0.4 0.2 25.3% 20.9% 10.1% 12.3% 12.2% 4.3% 90 75 51 41 45 10 23.3% 19.4% 13.2% 10.7% 11.7% 2.7% 0.6 14.8% 51 12.6% 0.5 14.9% 74 19.0% 149 711 250 83 8.2% 39.3% 13.8% 4.6% 962 891 305 161 31.6% 29.3% 10.0% 5.3% 615 34.0% 725 23.8% 2444 14.9% 5387 30.6%
R.I. McDonald et al. / Forest Ecology and Management 255 (2008) 4021–4031 4027 Table 2 (Continued ) White pine forests (species) Unharvested BA Harvested relBA Density relDens P. strobus B. lenta Q. rubra Q. velutina 7078 199 1162 581 Other 4992 BA relBA Density relDens 43.0% 1.2% 7.1% 3.5% 3245 2574 1972 1780 18.4% 14.6% 11.2% 10.1% 30.3% 2644 15.0% In a stratum in a particular forest type, only the most dominant species are listed. Basal area (BA) values are in m2/ha. the total number of seedlings is greater on harvested sites, the relative abundance of Quercus seedlings does not change substantially. There are few oak saplings, and no increase (or decrease) in oak saplings, in harvested versus unharvested sites. In Pinus-dominated stands, there is no indication of a significant change in P. strobus abundance in the seedling or sapling layers in response to harvesting. Harvested sites (average stump BA: 19.8 m2/ha) have no more seedlings than not/light
b Harvard Forest, Harvard University, 324 North Main Street, Petersham, MA 01366, United States 1. Introduction . Forest Ecology and Management 255 (2008) 4021-4031 ARTICLE INFO Article history: . significant change in vascular plant composition, non-metric multidimensional scaling revealed that
(A) boreal forest º temperate forest º tropical rain forest º tundra (B) boreal forest º temperate forest º tundra º tropical rain forest (C) tundra º boreal forest º temperate forest º tropical rain forest (D) tundra º boreal forest º tropical rain forest º temperate forest 22. Based on the
population ecology) and then subsequently covering interactions between species in a community (i.e., community ecology). However, to facilitate completion of the final paper, I have recently switched to covering community ecology and ecosystem ecology before population ecology. As both ecology and evolution have to be covered in the same .
Life science graduate education at Harvard is comprised of 14 Ph.D. programs of study across four Harvard faculties—Harvard Faculty of Arts and Sciences, Harvard T. H. Chan School of Public Health, Harvard Medical School, and Harvard School of Dental Medicine. These 14 programs make up the Harvard Integrated Life Sciences (HILS).
D. Mixed Evergreen/Deciduous Forest 38 1. Salt Dome Hardwood Forest * 38 2. Coastal Live Oak-Hackberry Forest * 39 3. Barrier Island Live Oak Forest * 39 4. Shortleaf Pine/Oak-Hickory Forest * 39 5. Mixed Hardwood-Loblolly Forest * 40 7. Slash Pine/Post Oak Forest * 40 8. Live Oak-Pine-Magnolia Forest * 40 9. Spruce Pine-Hardwood Flatwood * 41
Sciences at Harvard University Richard A. and Susan F. Smith Campus Center 1350 Massachusetts Avenue, Suite 350 Cambridge, MA 02138 617-495-5315 gsas.harvard.edu Office of Diversity and Minority Affairs minrec@fas.harvard.edu gsas.harvard.edu/diversity Office of Admissions and Financial Aid admiss@fas.harvard.edu gsas.harvard.edu/apply
reduce its forest road costs and still improve public safely on forest roads, reduce the impact of forest roads on the environment, and improve the ability of the Forest Service to fully maintain the national forest road system. For example, although the Forest Service already does some cost sharing, it could .
Submit questions to communityforest@fs.fed.us U.S. Forest Service Community Forest and Open Space Conservation Program Frequently Asked Questions (September 2013) General What is meant by the term "Community Forest"? The CFP rule defines Community Forest as "Forest land owned in fee-simple by an eligible entity that
Tank Gauge) API 2350 categorizes storage tanks by the extent to which personnel are in attendance during receiving operations. The overfill prevention methodology is based upon the tank catagory. Category 1 Fully Attended Personnel must always be on site during the receipt of product, must monitor the receipt continuously during the first and last hours, and must verify receipt each hour .
